The Effects of Mild Over-parameterization on the Optimization Landscape of Shallow ReLU Neural Networks
arXiv:2006.01005
Abstract
We study the effects of mild over-parameterization on the optimization landscape of a simple ReLU neural network of the form , in a well-studied teacher-student setting where the target values are generated by the same architecture, and when directly optimizing over the population squared loss with respect to Gaussian inputs. We prove that while the objective is strongly convex around the global minima when the teacher and student networks possess the same number of neurons, it is not even \emph{locally convex} after any amount of over-parameterization. Moreover, related desirable properties (e.g., one-point strong convexity and the Polyak-Łojasiewicz condition) also do not hold even locally. On the other hand, we establish that the objective remains one-point strongly convex in \emph{most} directions (suitably defined), and show an optimization guarantee under this property. For the non-global minima, we prove that adding even just a single neuron will turn a non-global minimum into a saddle point. This holds under some technical conditions which we validate empirically. These results provide a possible explanation for why recovering a global minimum becomes significantly easier when we over-parameterize, even if the amount of over-parameterization is very moderate.
References in corpus (8)
- Recovery Guarantees for One-hidden-layer Neural Networks
- Can SGD Learn Recurrent Neural Networks with Provable Generalization?
- Learning a Single Neuron with Gradient Methods
- Learning Parities with Neural Networks
- Approximate is Good Enough: Probabilistic Variants of Dimensional and Margin Complexity
- Learning Over-Parametrized Two-Layer ReLU Neural Networks beyond NTK
- On the Principle of Least Symmetry Breaking in Shallow ReLU Models
- Analytic Characterization of the Hessian in Shallow ReLU Models: A Tale of Symmetry